Page 60 - Simplicity is Key in CRT

60
Follow-up.
Clinical follow-up of patients consisted of physical examination, ECG and echocardiogram performed at least every 6 months. Follow-up of the device was performed at 1 and 3 months after CRT implantation and every 6 months thereafter. Data on hospitalization for acute heart failure were systematically collected. The diagnosis of heart failure required symptoms and signs consistent with congestive heart failure that was responsive to intravenous decongestive therapy.
Statistical analysis.
Continuous data are presented as median and 25th-75th percentiles (IQR) and categorical data as counts and percentages. Comparisons between groups were performed with Mann-Whitney U and Fisher’s exact tests, respectively for these data. Each patient’s ECG was classified as LBBB or not according to the five classification methods described above. Therefore, each patient could have from 0 to 5 LBBB definitions satisfied. Comparisons according to the number of LBBB positive classification per patient were performed using the Kruskall-Wallis test for continuous variables and the Fisher’s exact test for categorical variables. The test for trend was also applied. Significance was set at 0.003 for pairwise post-hoc comparisons.
The association of LBBB and reverse remodelling was assessed using a logistic regression model. Odds ratios (OR) and 95% confidence intervals (95%CI) were computed. Median follow-up (IQR) was computed with the inverse Kaplan Meier method. Event rates per 100 person year and 95%CI were computed.
Event-free survival was estimated by Kaplan-Meier method and compared with the logrank test. Hazard ratios (HR) and 95%CI were calculated with a Cox regression model. The proportional hazard assumption was satisfied in all cases. Endpoints for these analyses were HF hospitalization, all-cause mortality and the combination of these clinical endpoints. For both modelling procedures, both univariable and multivariable models with adjustment for a priori selected clinical confounders were fitted. Harrell’s c concordance statistic was computed for all Cox regression models including in turn each LBBB classification (the higher the Harrell’s c, the better model discrimination); classifications were informally compared by ranking the Harrell’s c. Statistical analyses were conducted using the Stata 15.1 software (Stata Corporation, College Station, TX, USA). A 2-sided p-value less than 0.05 was considered statistically significant.
Results
From January 2006 to December 2016, 498 patients received a CRT-device at the two participating institutions. Of those patients, 45 were excluded because of poor quality baseline ECG, 15 patients because of baseline paced QRS, 25 patients because of biventricular pacing <95%, and 97 patients were excluded due to incomplete follow- up data. Three-hundred sixteen patients were finally included in the analysis (Lugano n: 156; Maastricht n: 160). Demographic characteristics of the study cohort are summarized in Table 2.